Structural design using multi-objective metaheuristics. Comparative study and application to a real-world problem

Many structural design problems in the field of civil engineering are naturally multi-criteria, i.e., they have several conflicting objectives that have to be optimized simultaneously. An example is when we aim to reduce the weight of a structure while enhancing its robustness. There is no a single solution to these types of problems, but rather a set of designs representing trade-offs among the conflicting objectives. This paper focuses on the application of multi-objective metaheuristics to solve two variants of a real-world structural design problem. The goal is to compare a representative set of state-of-the-art multi-objective metaheuristic algorithms aiming to provide civil engineers with hints as to what optimization techniques to use when facing similar problems as those selected in the study presented in this paper. Accordingly, our study reveals that MOCell, a cellular genetic algorithm, provides the best overall performance, while NSGA-II, the de facto standard multi-objective metaheuristic technique, also demonstrates a competitive behavior.     

A limit set stabilization by means of the Port Hamiltonian system approach

A solution to the stabilization problem of a compact set by means of the Interconnection and Damping Assignment Passivity‐Based Control methodology, for an affine nonlinear system, was introduced. To this end, we expressed the closed‐loop system as a Port Hamiltonian system, having the property of almost all their trajectories asymptotically converge to a convenient limit set, except for a set of measure zero. It was carried out by solving a partial differential equation (PDE) or single matching condition, which allows the desired energy level or limit set E to be shaped explicitly. The control strategy was tested using the magnetic beam balance system and the pendulum actuated by a direct current motor (DC‐motor), having obtained satisfactory results. Copyright © 2014 John Wiley & Sons, Ltd.     

Real time motion estimation using a neural architecture implemented on GPUs

This work describes a neural network based architecture that represents and estimates object motion in videos. This architecture addresses multiple computer vision tasks such as image segmentation, object representation or characterization, motion analysis and tracking. The use of a neural network architecture allows for the simultaneous estimation of global and local motion and the representation of deformable objects. This architecture also avoids the problem of finding corresponding features while tracking moving objects. Due to the parallel nature of neural networks, the architecture has been implemented on GPUs that allows the system to meet a set of requirements such as: time constraints management, robustness, high processing speed and re-configurability. Experiments are presented that demonstrate the validity of our architecture to solve problems of mobile agents tracking and motion analysis.     

Providing security and fault tolerance in P2P connections between clouds for mHealth services

The mobile health (mHealth) and electronic health (eHealth) systems are useful to maintain a correct administration of health information and services. However, it is mandatory to ensure a secure data transmission and in case of a node failure, the system should not fall down. This fact is important because several vital systems could depend on this infrastructure. On the other hand, a cloud does not have infinite computational and storage resources in its infrastructure or would not provide all type of services. For this reason, it is important to establish an interrelation between clouds using communication protocols in order to provide scalability, efficiency, higher service availability and flexibility which allow the use of services, computing and storage resources of other clouds. In this paper, we propose the architecture and its secure protocol that allows exchanging information, data, services, computing and storage resources between all interconnected mHealth clouds. The system is based on a hierarchic architecture of two layers composed by nodes with different roles. The routing algorithm used to establish the connectivity between the nodes is the shortest path first (SPF), but it can be easily changed by any other one. Our architecture is highly scalable and allows adding new nodes and mHealth clouds easily, while it tries to maintain the load of the cloud balanced. Our protocol design includes node discovery, authentication and fault tolerance. We show the protocol operation and the secure system design. Finally we provide the performance results in a controlled test bench.     

Advantages of Model Driven Engineering for studying complex systems

The evaluation of the emergent behaviour in complex systems requires an analytical framework which allows the observation of different phenomena that take place at different levels. In order to observe the dynamics of complex systems, it is necessary to perform simulations so that both local and the emergent behaviour can be observed. To this end, the way in which complex system simulators are built must be examined so that it will be feasible to model large scale scenarios. In this paper, the use of Model Driven Engineering methodology is proposed to deal with this issue. Among other benefits, it is shown that this methodology allows the representation and simulation of a complex system providing support for the analysis. This analysis is supported by a metamodel which describes the system components that are under study. The application of this methodology to the development of large scale simulators is explored through a case study. This case study analyses a complex socio-technical system: a power grid.     

Multi-objective optimization with an adaptive resonance theory-based estimation of distribution algorithm

The introduction of learning to the search mechanisms of optimization algorithms has been nominated as one of the viable approaches when dealing with complex optimization problems, in particular with multi-objective ones. One of the forms of carrying out this hybridization process is by using multi-objective optimization estimation of distribution algorithms (MOEDAs). However, it has been pointed out that current MOEDAs have an intrinsic shortcoming in their model-building algorithms that hamper their performance. In this work, we put forward the argument that error-based learning, the class of learning most commonly used in MOEDAs is responsible for current MOEDA underachievement. We present adaptive resonance theory (ART) as a suitable learning paradigm alternative and present a novel algorithm called multi-objective ART-based EDA (MARTEDA) that uses a Gaussian ART neural network for model-building and a hypervolume-based selector as described for the HypE algorithm. In order to assert the improvement obtained by combining two cutting-edge approaches to optimization an extensive set of experiments are carried out. These experiments also test the scalability of MARTEDA as the number of objective functions increases.